ChatGPT can't see your diary

Buried in this year's Private Practice Barometer, past the salary tables, is the most useful finding in the whole report, and it isn't the adoption number. It's the shape of the reluctance. Only 14% of clinic owners are actually resistant to AI. Only 8% say cost is the blocker. The two barriers that top the list, at 18% apiece: 'I'm unsure which tools are actually useful' and 'I don't know where to start.'

That isn't resistance. That's a queue of people standing in the aisle, wallet out, with no idea what to put in the basket.

Now look at what's actually in the basket. Asked what AI they use, owners named ChatGPT more than every other tool combined. Asked what they most want automated, they said three things: collecting reviews, reactivating lapsed patients, marketing. Hold those two lists next to each other for a moment. They barely touch.

A chatbot will happily draft the review request. What it cannot do is know which patient finished a course of care on Tuesday feeling brilliant, which one quietly lapsed after session three, and which one was properly discharged and should be left alone. It can't, because it can't see your diary. The general-purpose tools everyone bought are eloquent about everything and informed about nothing that actually runs a clinic.

The most-wanted automation in the industry is the one a chatbot is least able to do.

There's a clue in the survey about how this resolves, and it's the automation nobody asked for. Appointment reminders barely registered on the wishlist. Not because owners don't value them, but because the problem is already solved. The PMS sends them. Silently, reliably, without anyone opening a tab. Reminders stopped being 'AI' the day they disappeared into the workflow. That's the pattern for everything else on the list.

Clinicians already have a test for this, the same one you'd apply to any new modality that turns up promising outcomes: does it change the treatment plan, or is it a gadget? A tool that reads the diary, knows who's lapsed and who's discharged, and does the follow-up at the right moment changes the plan. A tab you have to remember to open, then copy things in and out of, is a gadget with good manners.

This is, transparently, the shape of thing we build, so weigh my view accordingly. But the survey's own commentary lands near the same spot: owners aren't waiting for cheaper AI, they're waiting to be told which tools to trust. I'd put it differently. Stop auditioning chatbots. Ask one question of anything you're offered: can it see the diary? If it can't, it isn't automating your clinic. It's automating your typing.

The least confident group in the data, tellingly, is the smallest clinics, the ones an extra pair of invisible hands would help most. Which makes the confusion the expensive part, not the tools. Most of the industry is willing. The aisle just needs fewer gadgets on the shelf.